Fixing Up Typography - Page 3

ahi · 05-17-2009, 03:39 PM

Some of the comments, and particularly the special focus on quotation marks makes me wonder if my original "dream" of a single script to take a plaintext document from A to Z is misguided. (Even though, as I stated earlier, it seems to me GutenMark does a reasonable job of that.)

Perhaps a better way would be to write small utilities each of which focus on just one aspect of the document cleanup/conversion/fix-up process. I myself might play around with a quotation mark fixing utility, when I get a chance over the next week or so.

Some other utilities I can think of:

- metadata recognizer (i.e.: figures out title, author, chapter titles, et al)
- paragraph normalizer (remove manual linebreaks between lines, keep only one between paragraphs)
- emphasis normalizer (convert the myriad ways of indicating emphasis into a single standard [and ideally simple to accurately parse] markup)

All of these utilities I think of as being command line tools that do as much as possible without human intervention but ask for human/manual arbitration when up against a case that requires a judgement call (or, rather, actually understanding the text).

Does anything like this exist? Is the idea kind of crazy, or kind of sensible?

Sincerely,

AHI

pepak · 05-17-2009, 04:01 PM

It is sensible, I am just a bit doubtful about it being realistic :-)

I was thinking along these lines, but really, even with same-source files, there are too many variations so that it's too much effort writing a fully-automatic utility. I just keep a collection of useful regexps (or normal replaces) and in each case decide which one (and in which order) to use.

ahi · 05-17-2009, 04:07 PM

Quote:

Originally Posted by pepak

It is sensible, I am just a bit doubtful about it being realistic :-)

I was thinking along these lines, but really, even with same-source files, there are too many variations so that it's too much effort writing a fully-automatic utility. I just keep a collection of useful regexps (or normal replaces) and in each case decide which one (and in which order) to use.

Well, in a way, a utility could be as simple as a set of regexes, with command-line switches (or heuristics at start-up time) mandating which specific one to use. I do think though that some of the fix-ups that can be addressed using regexes might be addressed more reliable through simpler sequential processing. (Though, as with all things, I could be wrong.)

It definitely is a tall order... but broken down into its components like this might make it more achievable. Like I said, I might try my hands at the quotation stuff in the near future.

Jellby · 05-18-2009, 05:11 AM

Quote:

Originally Posted by rogue_ronin

Is ' well-supported now?

Unfortunately (as far as I know) there is no different character for a curly apostrophe, we have to use a single right quote (' is a straight apostrophe, just as " is a straight double quote). That's why I use ’ and & #8217; for single right quotes and apostrophes, they are exactly the same character, with two different names, but at least I can keep them different in the source file.

rogue_ronin · 05-18-2009, 05:20 AM

Hmmm, but that's for presentation, right? Is there a way to hack that with CSS? Calling an apostrophe a quote makes for inaccurate meta.

(Okay, now that I've begun thinking about it, CSS is becoming like a magical pony, bringing gifts and hopes to all the children.)

m a r

Sweetpea · 05-18-2009, 06:01 AM

Quote:

Originally Posted by rogue_ronin

I think I need to study up on it. I'm using a sort of bastardized HTML mix. I've been using clips (macros) in NoteTab for a long time now, so I re-checked what I'm doing. I actually use both <a name="chapter_ChapterNumber"> and <h3 id="chapter_ChapterNumber" class="chapter" align="center"> in my files to mark a chapter. A combination of overkill and ignorance.

I used to use that too. Untill I started making my source files epub compliant. The <a name=""> isn't valid with that. So, I only use the id now.

Jellby · 05-18-2009, 06:11 AM

Quote:

Originally Posted by rogue_ronin

Hmmm, but that's for presentation, right? Is there a way to hack that with CSS? Calling an apostrophe a quote makes for inaccurate meta.

I agree that's somewhat undesirable... maybe another solution would be keeping the ' in the source and postprocessing it to a curly apostrophe when creating a particular end format, but I don't think that can be done with CSS (at least not with the CSS2 used in ePUB).

rogue_ronin · 05-18-2009, 08:02 AM

Can you name your own entities with CSS2?

@sweetpea: yeah, I'm starting to get that...

m a r

Jellby · 05-18-2009, 10:41 AM

Quote:

Originally Posted by rogue_ronin

Can you name your own entities with CSS2?

Hm... good question.

I don't think you can with CSS2, but maybe there are other means. See an example here (view the source). Now the question is whether this can be used with ePUB.

DaleDe · 05-18-2009, 12:53 PM

Quote:

Originally Posted by HarryT

I gave up on "curly quotes" a long time ago, for that very reason. If the source has them, fine. If it doesn't, it's a virtually impossible task to programatically get them right; you certainly can't assume that they always occur in pairs!

I find that word does a pretty good job of handling curly quotes. It may not be perfect but does the right thing the majority of the time. You just turn on smart quotes and then replace " with " and the magic happens. Do it again with ' with '.

Dale

DaleDe · 05-18-2009, 01:00 PM

Quote:

Originally Posted by Jellby

Hm... good question.

I don't think you can with CSS2, but maybe there are other means. See an example here (view the source). Now the question is whether this can be used with ePUB.

I have seen this done in ePUB but not all readers seem to support it. Generally it is add at the top of the xhtml file, not in the CSS file.

Dale

pepak · 05-18-2009, 03:36 PM

Quote:

Originally Posted by rogue_ronin

Can you name your own entities with CSS2?

No, but you can name your own entities with XHTML :-)

rogue_ronin · 05-18-2009, 11:56 PM

So that means you could re-name them as well, right?

Such that for problems like $apos; -- you could use the common entity-name, then anyone that wanted to could redefine it as a $rsquo; or whatever they like.

Still, it'd be better if it could be called externally to the ebook file, since it's really a presentation issue, no?

m a r

Jellby · 05-19-2009, 06:08 AM

I don't know if you can actually redefine ', since it's one of the predefined XML entities, but it looks like you can redefine — and others. In any case, it's possible to create a new ≈ entity and make it look whatever you like. I'll do some tests with this...

rogue_ronin · 05-19-2009, 07:43 AM

I was just looking at the CSS tutorial at w3schools.

I'm wondering if a different <link> tag in the head might allow you to call some customizations? I'm just spit-balling here -- I have no idea how the <link> tag works; but if the CSS file is read in the head, it makes me wonder if you could use one of the other rel="" attribute/values to call a set of entity redefinitions? Like maybe a "section" or a "bookmark"

Anyway, I'll leave it to you experts.

m a r

05-19-2009, 07:43 AM	#45
rogue_ronin Banned Posts: 475 Karma: 796 Join Date: Sep 2008 Location: Honolulu Device: Nokia 770 (fbreader)	I was just looking at the CSS tutorial at w3schools. I'm wondering if a different <link> tag in the head might allow you to call some customizations? I'm just spit-balling here -- I have no idea how the <link> tag works; but if the CSS file is read in the head, it makes me wonder if you could use one of the other rel="" attribute/values to call a set of entity redefinitions? Like maybe a "section" or a "bookmark" Anyway, I'll leave it to you experts. m a r Last edited by rogue_ronin; 05-19-2009 at 07:44 AM. Reason: Add Link

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Kindle Typography	ChaoZ	Amazon Kindle	21	08-14-2010 12:50 PM
Is there hope for better ebook typography?	tomsem	Amazon Kindle	0	08-12-2010 10:44 PM
Typography on the iPad	LDBoblo	Apple Devices	1	04-14-2010 03:33 PM
French Typography	ahi	Workshop	14	09-16-2009 02:22 PM
Chinese Typography	ahi	Workshop	81	09-14-2009 09:34 AM

05-17-2009, 03:39 PM	#31
ahi Wizard Posts: 1,790 Karma: 507333 Join Date: May 2009 Device: none	Some of the comments, and particularly the special focus on quotation marks makes me wonder if my original "dream" of a single script to take a plaintext document from A to Z is misguided. (Even though, as I stated earlier, it seems to me GutenMark does a reasonable job of that.) Perhaps a better way would be to write small utilities each of which focus on just one aspect of the document cleanup/conversion/fix-up process. I myself might play around with a quotation mark fixing utility, when I get a chance over the next week or so. Some other utilities I can think of: - metadata recognizer (i.e.: figures out title, author, chapter titles, et al) - paragraph normalizer (remove manual linebreaks between lines, keep only one between paragraphs) - emphasis normalizer (convert the myriad ways of indicating emphasis into a single standard [and ideally simple to accurately parse] markup) All of these utilities I think of as being command line tools that do as much as possible without human intervention but ask for human/manual arbitration when up against a case that requires a judgement call (or, rather, actually understanding the text). Does anything like this exist? Is the idea kind of crazy, or kind of sensible? Sincerely, AHI

05-17-2009, 04:01 PM	#32
pepak Guru Posts: 610 Karma: 4150 Join Date: Mar 2008 Device: Sony Reader PRS-T3, Kobo Libra H2O	It is sensible, I am just a bit doubtful about it being realistic :-) I was thinking along these lines, but really, even with same-source files, there are too many variations so that it's too much effort writing a fully-automatic utility. I just keep a collection of useful regexps (or normal replaces) and in each case decide which one (and in which order) to use.

05-18-2009, 05:20 AM	#35
rogue_ronin Banned Posts: 475 Karma: 796 Join Date: Sep 2008 Location: Honolulu Device: Nokia 770 (fbreader)	Hmmm, but that's for presentation, right? Is there a way to hack that with CSS? Calling an apostrophe a quote makes for inaccurate meta. (Okay, now that I've begun thinking about it, CSS is becoming like a magical pony, bringing gifts and hopes to all the children.) m a r

05-18-2009, 08:02 AM	#38
rogue_ronin Banned Posts: 475 Karma: 796 Join Date: Sep 2008 Location: Honolulu Device: Nokia 770 (fbreader)	Can you name your own entities with CSS2? @sweetpea: yeah, I'm starting to get that... m a r

05-18-2009, 11:56 PM	#43
rogue_ronin Banned Posts: 475 Karma: 796 Join Date: Sep 2008 Location: Honolulu Device: Nokia 770 (fbreader)	So that means you could re-name them as well, right? Such that for problems like $apos; -- you could use the common entity-name, then anyone that wanted to could redefine it as a $rsquo; or whatever they like. Still, it'd be better if it could be called externally to the ebook file, since it's really a presentation issue, no? m a r

05-19-2009, 06:08 AM	#44
Jellby frumious Bandersnatch Posts: 7,515 Karma: 18512745 Join Date: Jan 2008 Location: Spaniard in Sweden Device: Cybook Orizon, Kobo Aura	I don't know if you can actually redefine ', since it's one of the predefined XML entities, but it looks like you can redefine — and others. In any case, it's possible to create a new ≈ entity and make it look whatever you like. I'll do some tests with this...

Advert

Advert